Automated Control and Data Analysis for Multicellular Multidetector Microscale Gas Chromatographs
Xu, Qu
2024
Abstract
Industrial manufacturing processes often release volatile organic compounds (VOCs). VOCs can be harmful to humans and the environment but can also serve as markers for process control. The timely detection and recognition of VOCs are, therefore, important. Gas chromatography (GC) systems are the current gold standard for VOC detection, but are limited by their high bulk, cost, and power consumption. Microscale gas chromatographs (μGCs) are miniaturized GC instruments that can be distributed and deployed more easily, thus show greater promise for automated in situ analysis. As μGCs become increasingly sophisticated, their controls and automated recognition become increasingly complex as well. Current μGC systems often require trained personnel to operate. Automation of controls and VOC recognition is, therefore, necessary to achieve fully autonomous and automated in situ analysis. This work investigates the challenges of automating the controls, and recognition of VOCs in highly complex μGC systems. The representative μGC system is based on a multicellular progressive cellular architecture that uses multiple cells to broaden the range of analytes. The multiple detectors within each cell further enhance chemical recognition. The control software manages data acquisition of the μGC system in a time sensitive manner, while operating multiple control loops and error conditions. The multithreaded control software enables concurrent control of heaters, pumps, and valves, while also gathering data from thermistors, pressure sensors, capacitive detectors, and photoionization detectors. A graphical user interface (UI), implemented on a laptop computer, provides remote control and real time data visualization. In experimental evaluations, the control software provided successful automation of all the components, including 8 sets of thermistors and heaters for temperature feedback loops, 2 sets of pressure sensors and pumps that form pressure head feedback loops, 6 capacitive detectors, 3 photoionization detectors, 6 valves, and a fixed-flow gas pump. A typical run analyzing 18 chemicals is presented. Despite the use of a non-real-time operating system, the standard deviations of the control loop timings were <0.5% of the intended time interval between measurements. The control software successfully supported >1000 μGC runs that analyzed various mixtures. A chemical recognition algorithm can be a valuable part of any autonomous μGC system. For a multi-detector μGC system, the chemical analysis must account for the retention time of each chemical analyte as well as the relative response of each detector to each analyte, i.e., the detector response pattern (DRP). This work reports a rule-based automated chemical recognition algorithm for a multi-cell, multi-detector μGC system. The algorithm applies rules based on expert knowledge to compare the detected peaks. Consequently, this algorithm only requires a small amount of calibration data but not extensive training data. Additionally, the algorithm provides special subroutines for chromatogram peaks with a small signal-to-noise ratio and for asymmetrical peaks that may result from surface adsorptive analytes. The algorithm was verified by multiple experimental tests. Each test included chromatograms with 21–31 peaks for each detector. The true positive rate was 96.3%, the true negative rate was 94.1%, the false positive rate was 5.9%, and the false negative rate was 3.7%. The results demonstrated that the algorithm could support μGC systems for automated chemical screening and early warning applications.Deep Blue DOI
Subjects
microscale gas chromatography MEMS Control Software Expert System Firmware
Types
Thesis
Metadata
Show full item recordCollections
Remediation of Harmful Language
The University of Michigan Library aims to describe library materials in a way that respects the people and communities who create, use, and are represented in our collections. Report harmful or offensive language in catalog records, finding aids, or elsewhere in our collections anonymously through our metadata feedback form. More information at Remediation of Harmful Language.
Accessibility
If you are unable to use this file in its current format, please select the Contact Us link and we can modify it to make it more accessible to you.